ggerganov/llama.cpp

Pull Requests Commits

llama : respect no_alloc in all memory caches (#31087)

ggerganov committed 3 days ago

Verified fa3645fe

tools : reject truncated GGUF tensor data in merge (#31044)

shreya-11733 committed 3 days ago

Verified 7790a923

CUDA: fuse FFN gate/up matmuls and GLU in MMQ with interleaved gate/up rows (#29948)

anujj committed 3 days ago

Verified c49edf09

ggml-cuda: fuse qwen MRoPE + set-rows (#31025)

am17an committed 3 days ago

Verified a55c16e4

vulkan: Add a larger tile size for conv3d (coopmat2 only) (#29920)

jeffbolznv committed 3 days ago

Verified e321bdad

vulkan : one warp tile per subgroup in the small matmul tiles (#30106)

ianloic committed 3 days ago

Verified 459a1ba5

hexagon: generic dims collapse/reshape for unary, binary and matmul ops (#31071)

max-krasnyansky committed 3 days ago

Verified f0fb5984

opencl: add topk and lightning indexer (#31026)

lhez committed 3 days ago

Verified 9a4580b6

opencl: add initial q2_k support (#30999)

lhez committed 3 days ago

Verified b083c384

ci : update CPU job assignments (#31136)

ggerganov committed 3 days ago

Verified 344be0a6

chat : reject JSON nested more than 128 levels before walking it (#30242)

ykhrustalev committed 4 days ago

Verified b3fa6995

model: DSV4.1 flash (#31086)

am17an committed 4 days ago

Verified 74bca395

jinja : fix select and selectattr argument handling (#31059)

JoeyPatricio committed 4 days ago

Verified d327bf1f

hexagon: KV-head team split for multicore HMX flash attention (#31055)

ebateni committed 4 days ago

Verified 78a8ae14

server: refactor server entrypoint (#31113)

ngxson committed 4 days ago

Verified f3269ef9

metal : remove mul_mm_id amax rescaling (#30924)

ggerganov committed 4 days ago

Verified 3f03b177

rpc : add bounds check in rdma_recv (#30399)

sohaibharraoui committed 4 days ago

Verified 9a0fde99

ui: Models Manager as Route + UI improvements (#30988)

allozaur committed 4 days ago

Verified e168e041

fix(ui): align hover card and context popup styling with dropdown menu (#30978)

allozaur committed 4 days ago

Verified ea1d7d66

ci : pass per-runner results dir to self-hosted jobs on T4 runners (#31100)

ggerganov committed 4 days ago

Verified 71255fef

minimax-m3 : forward quant scales to the graph (fixes #31063) (#31097)

origin-labs-ai committed 4 days ago

Verified 73adf89e

tests : fix test-llama-context for large n_batch (#31083)

sanmai committed 4 days ago

Verified df9beac7

ci : exclude test-speculative-rejection-sampling from macos-latest-arm64 (#31088)

ggerganov committed 4 days ago

Verified 4d32b29a

llama : deny -sm tensor for arches with moe amax rescaling (#31081)

ggerganov committed 4 days ago

Verified 30322097

ggml : allow MUL_MAT_ID -1 index to skip the computation (#26631)

ngxson committed 4 days ago

Verified 701fd25b

tests : add speculative rejection draft sampling test (#30213)

danbev committed 4 days ago

Verified 3642238c

ci : fix data race between runners sharing the model cache (#31046)

ggerganov committed 4 days ago

Verified 6b9ba4f2

vulkan: ggml src1 precision support (#30656)

0cc4m committed 5 days ago

Verified 8e5a58df

HIP: add byte permute fallback when __builtin_amdgcn_perm is unavailable (#30915)

aobolensk committed 5 days ago

Verified 1ddffe9f

mtmd: move build_inp_temporal_merge to base class (#31065)

ngxson committed 5 days ago

Verified 7894e8b9

Newer Older